> For the complete documentation index, see [llms.txt](https://stereotoolss-organization.gitbook.io/stomics-software-download/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://stereotoolss-organization.gitbook.io/stomics-software-download/add-on-packages/saw.md).

# SAW

## Plugin

Plugin module of SAW support is designed to **manage extension packages** that expand the capabilities of the SAW software. In version 8.3, we released the GPU aceleration package, a critical extension designed to significantly boost computational performance, especially for large Stereo-seq chip datasets.

### GPU acceleration package 1.0 (Aug., 2026)

* **Key feature:** accelerates downstream matrix analysis (PCA, UMAP, clustering, etc.)
* **Scope:** applies to the step of secondary analysis, **not to image processing**
* **Requirement**: SAW **8.3.0 or higher** required

{% hint style="info" %}
Before downloading, please verify your server's **CUDA version** (e.g., by running the `nvidia-smi` command) to ensure you select the compatible GPU acceleration package.
{% endhint %}

<table data-card-size="large" data-view="cards"><thead><tr><th></th><th></th></tr></thead><tbody><tr><td><a href="https://enfile.stomics.tech/saw-plugin-gpu-cu11-1.0.tar.gz"><strong>Download GPU package for CUDA 11 (tar.gz)</strong></a></td><td></td></tr><tr><td>File size: 6.5 GB</td><td>md5sum: 87a089d6160ddc1b357d79951c66d3d8</td></tr><tr><td><a href="https://enfile.stomics.tech/saw-plugin-gpu-cu12-1.0.tar.gz"><strong>Download GPU package for CUDA 12 (tar.gz)</strong></a></td><td></td></tr><tr><td>File size: 5.0 GB</td><td>md5sum: 47ca19b64520d0b7883536ee481885b9</td></tr><tr><td><a href="https://enfile.stomics.tech/saw-plugin-gpu-cu13-1.0.tar.gz"><strong>Download GPU package for CUDA 13 (tar.gz)</strong></a></td><td></td></tr><tr><td>File size: 2.5 GB</td><td>md5sum: 907255c236ae21de72a9b1ace9a668e4</td></tr></tbody></table>

{% tabs %}
{% tab title="wget" %}

```bash
#CUDA 11
wget -c "https://enfile.stomics.tech/saw-plugin-gpu-cu11-1.0.tar.gz"
#CUDA 12
wget -c "https://enfile.stomics.tech/saw-plugin-gpu-cu12-1.0.tar.gz"
#CUDA 13
wget -c "https://enfile.stomics.tech/saw-plugin-gpu-cu13-1.0.tar.gz"
```

{% endtab %}

{% tab title="curl" %}

```shellscript
#CUDA 11
curl -C - -O "https://enfile.stomics.tech/saw-plugin-gpu-cu11-1.0.tar.gz"
#CUDA 12
curl -C - -O "https://enfile.stomics.tech/saw-plugin-gpu-cu12-1.0.tar.gz"
#CUDA 13
curl -C - -O "https://enfile.stomics.tech/saw-plugin-gpu-cu13-1.0.tar.gz"
```

{% endtab %}
{% endtabs %}

By installing the GPU acceleration package, SAW can offload computationally intensive tasks to the GPU, drastically reducing runtime for the following analysis steps:

* Dimensionality reduction: accelerates PCA (Principal Component Analysis) and UMAP (Uniform Manifold Approximation and Projection).
* Clustering: speeds up Leiden clustering.
* Joint analysis: enhances the efficiency of multi-omics/multi-sample integration and batch effect correction (e.g., in `saw aggr`).